Introduction: Ants in tropical forests are a hyper-diverse group that plays important ecological roles. Several studies on tropical forests have used different sampling protocols to capture soil ants, making it difficult to compare responses and patterns of diversity between studies. Thus, research that compares different well-structured and standardized sampling methodologies to adequately estimate the richness of ant species in tropical forests is necessary. Objective: In this study, we examined the combination of catches with pitfall traps with and without baits and litter collections for soil ant sampling. Methods: In the fall traps, we use two baited (sardines and bananas) and one non-baited. For the manual collections (litter sampling), the litter and only the topsoil of the loose soil were collected. Results: We found that traps containing sardine baits collected a greater abundance of ants, whereas non-baited traps collected a greater richness of ant species. On the other hand, litter collections captured the largest number of exclusive species, presenting a different species composition from the pitfall traps (with and without baits). In general, baited traps showed greater abundance in more degraded locations, while manual collections and unbaited pitfalls captured more individuals in preserved environments. Conclusions: Our results provide evidence that, for accurate sampling of soil ant assemblages in tropical forests, the use of different methodologies is necessary to capture a greater diversity of species because the methods differ in effectiveness according to habitat.

Keywords: ant community, Formicidae, inventory, sampling methods, tropical forests